Did you know that two thermometers can give different answers even when both are placed in the same cup of water? One student reads 25.0 °C, another reads 25.8 °C, and the class wants to announce that one instrument is broken. It is a tempting conclusion. It is also too early.
Comparing disagreeing thermometers means checking whether both instruments measured the same material under the same conditions, calculating the observed difference, inspecting their resolution and behaviour, then using a trusted reference where appropriate. A difference establishes disagreement; by itself it does not identify which reading is true or which instrument is inaccurate. That is the immediate lesson and the starting point for an honest investigation.
First diagnose the measurement, not the pupil
Begin by asking what the students actually measured. One probe may have touched the side of a warm beaker. Another may have been suspended in the water. One thermometer may still be responding to a previous room temperature. Another may have been read from above rather than at eye level. Even a correctly calibrated thermometer can disagree with another if they were used in different places, at different moments or before the temperature became stable.
A class can learn more from a careful disagreement than from two matching numbers copied into an attractive table. The aim is not to make the instruments agree at all costs. It is to understand the conditions under which a reading becomes a fair comparison.
What a temperature measurement actually tells us
A thermometer reports a temperature indication, in this article measured in degrees Celsius (°C). To compare indications, the class must identify the relevant medium, the position and depth of each sensing part, the measurement time and the instrument scale. A digital probe with 0.1 °C display steps should not automatically be treated as accurate to 0.1 °C; display resolution and measurement accuracy are different properties.
At school level, students can use the terms resolution (the smallest displayed step), repeatability (how similar repeated results are under similar conditions), and accuracy (closeness to a suitable reference value). Those are separate questions. An instrument that repeatedly shows 25.7 °C may be consistent but still systematically offset from a trusted reference.
For deeper metrology, the US National Institute of Standards and Technology reports comparisons of thermometer calibrations and measurement uncertainties. The school exercise below does not claim the precision of those specialised laboratories; it borrows their useful principle that readings should be compared with awareness of conditions and uncertainty.
Worked example: two stable instruments, one cup
Fictional classroom data for practice: a teacher prepares a well-mixed water bath within the safe operating range of two suitable school digital thermometers. Both probes are placed at comparable depths without touching the container. The display values are allowed to settle. Three paired observations, each made while the bath is checked to be broadly stable, are recorded.
- Observation 1: Thermometer A = 24.8 °C; Thermometer B = 25.5 °C.
- Observation 2: A = 25.0 °C; B = 25.7 °C.
- Observation 3: A = 24.9 °C; B = 25.6 °C.
- Mean of A = (24.8 + 25.0 + 24.9) ÷ 3 = 24.9 °C.
- Mean of B = (25.5 + 25.7 + 25.6) ÷ 3 = 25.6 °C.
- Difference between the two means = 25.6 − 24.9 = 0.7 °C.
Each fictional pair differs by 0.7 °C, so the observed gap is consistent across these three readings. The evidence supports the statement: ‘Thermometer B indicated 0.7 °C more than A under these recorded conditions.’ It does not support ‘B is 0.7 °C too high’. A could be low, B could be high, both could be offset, or other systematic influences could still be present.
If both devices display only tenths of a degree, copying extra digits into the conclusion would imply unsupported precision. Means are calculations on the recorded readings, not a magical upgrade to the accuracy of the instruments. Where uncertainty is relevant, discuss it explicitly rather than reporting more decimals merely because a calculator can.
The important distinction: disagreement, random variation and bias
Disagreement
The simple fact that two indications differ. Students can calculate it directly without an external reference. If one reads 25.0 °C and the other 25.8 °C, the observed difference is 0.8 °C. The size of the gap is a fact about the pair, not yet a diagnosis of cause.
Variation across repeated observations
A thermometer may fluctuate between displayed steps or respond to a genuinely changing water temperature. Repeating the same flawed setup does not remove every systematic problem. Record conditions alongside repeats and do not treat repeated readings as independent experiments unless the apparatus was meaningfully reset.
Systematic offset
If one instrument repeatedly reads higher, it might have a consistent offset under those conditions. Demonstrating the error relative to true temperature requires an appropriate reference or calibrated method. Do not take the lower reading to be correct by default, and do not use ‘the average of two’ as unquestionable truth.
A simple school comparison protocol
- Read the device instructions and check that both are suitable for the temperature range and measurement medium. The teacher selects any reference method and supervises the procedure.
- Record the make or identifying label, units and display step. If one instrument uses °F, convert units carefully before comparing; do not subtract a Fahrenheit value from a Celsius value.
- Place sensors at comparable positions and immersion depths. Avoid contact with the vessel’s base or side unless the instrument instructions specifically call for that arrangement.
- Wait until the readings have settled sufficiently for the task. Record times and avoid taking one reading in the warming phase and the other after cooling begins.
- Write each displayed number exactly as seen. Do not secretly adjust a value to make it match the other thermometer.
- Repeat, compare the differences, and, where available, check against an appropriate trusted reference rather than proclaiming a winner from agreement alone.
Do not ask pupils to test instruments using boiling water, open flames or electrical improvisations. Hot-water work and calibration checks require competent adult control. A simple ambient or teacher-prepared safe water-bath comparison is enough to teach the principle.
How to work out whether there is a shared cause
Suppose a classroom cup sits near a sunny window. Both thermometers may steadily rise as the water warms, so even excellent instruments could show a changing temperature. A class that takes the A reading at 10:00 and the B reading at 10:05 is not really testing disagreement alone; it has mixed instrument differences with changing conditions.
A better arrangement reads both instruments at approximately the same time, records a common observation time and checks whether the bath is changing. If practical and safe, the teacher can exchange probe positions to see whether a difference follows a particular instrument or a particular place in the water. That manoeuvre is diagnostic, not automatic proof of calibration error.
An independent check with a third reference
Imagine a certified reference device, suitable for the range and properly used, indicates 25.2 °C with a stated uncertainty that the teacher understands. If A shows 24.9 °C and B shows 25.6 °C at the same time and location, the deviations relative to the reference indication are −0.3 °C and +0.4 °C, respectively. These are differences from the reference indication, not statements of exact errors in an unknown true temperature. Uncertainty and possible differences in sensor positioning still matter.
If no credible reference is available, students should report the disagreement and their controls, not invent a corrected temperature. The honest observation is a successful learning outcome, even when the class cannot completely resolve the cause.
Learner diagnostic: which statement is defensible?
- Two thermometers read 21.5 °C and 22.1 °C at the same moment. ‘They differ by 0.6 °C’ is defensible. ‘The first is wrong’ is not yet established.
- Three consecutive readings of one probe are 20.0 °C, 20.0 °C and 20.0 °C. Does this prove accuracy? No. It demonstrates consistency at the displayed step under those conditions.
- A thermometer was moved from cool air into warmer water and its numbers are still rising. Is the first displayed number the stable water temperature? Not necessarily; response time is part of the method.
- An instrument reads 73 °F and another 23 °C. Can the student subtract 73 − 23? Not meaningfully until both readings are expressed in the same temperature scale.
Progressive practice: teach the reasoning in stages
Guided: calculate the disagreement
A reads 19.4 °C and B reads 19.9 °C. Subtract 19.9 − 19.4 = 0.5 °C. Explain why the result has temperature-difference units and does not assign blame to either instrument.
Supported: identify one variable to repair
Two thermometers are dipped into a tall beaker, but one sensor rests on the bottom and the other is halfway up. Pupils design a repeat with comparable sensor positions, record the change and describe whether the disagreement becomes smaller, larger or unchanged. They must explain why a single improved agreement does not certify accuracy.
Independent: write a defensible conclusion
Give students a fictional dataset A = 18.3, 18.4, 18.5 °C and B = 18.9, 19.0, 19.1 °C. They compute mean A = 18.4 °C, mean B = 19.0 °C and mean difference = 0.6 °C. Then ask them to list two plausible unresolved causes and a safe next step. The answer needs a calculation and an evidence limit.
A useful teacher and parent observation checklist
Look beyond whether the student obtained 0.6 or 0.7. Can the student name the controlled conditions, preserve units, distinguish agreement from truth and refuse an unsupported claim? A parent can practise with two ordinary room thermometers in the same room without introducing hot water. The goal is to ask what was held constant, not to demand identical numbers.
Track progress with two short tasks completed on different days: first, compute the gap from already collected readings; second, propose a fair protocol and justify a cautious conclusion. A learner who can independently explain why a reference matters has learned something durable beyond the worksheet.
Limits and frequently asked questions
Should I just average the two readings?
An average can describe the pair, but it does not necessarily estimate the true temperature correctly if one or both instruments have systematic errors. Explain why the average is being used and what it can and cannot establish.
Does a digital display mean a thermometer is accurate?
No. Digital describes how information is displayed. Check stated performance, suitable range, procedure and calibration if the decision requires it.
What if the water itself changes temperature?
Record observation times and consider whether the medium was stable or well mixed. Do not interpret a changing environment as a pure difference between devices.
Does this exercise prove that one brand is better?
No. A few classroom readings under one setup cannot justify broad product comparisons. Different instruments may be designed for different ranges, response speeds and applications.
Where to learn more
For related skills, read accuracy versus precision and when a temperature probe has settled. The eduKate Education Hub gives a wider learning route, while the Science learning guide and the Clementi diagnosis-first reference illustrate how explanation and checking should lead practice.
The core aim here is straightforward: when two instruments disagree, record what they actually say, improve the comparison, and claim only what the evidence can carry. That habit is more valuable than a quick guess about which thermometer to trust.